Bird Strike Committee Proceedings


Date of this Version

May 1999


Certification of aircraft against the birdstrike threat is expensive and time consuming. With the need to reduce design life cycle time and costs with ever more complex structures (materials, geometries and manufacturing methods) yet with no reduction in safety, the possibilities of certification via generic analysis is an attractive proposition. This paper discusses one approach being considered within British Aerospace (BAe). It relies heavily on research activities that have derived extensive data for bird biometrics and innovative testing that can provide mechanical data for bird failure modes unique to military aircraft.

Whilst almost all birdstrike clearance is performed via testing using real birds on representative structure, some alternatives methods are now possible. Reference 1 for civil aircraft, states that "compliance may be shown by analysis" but that the analysis must be based on tests performed on "sufficiently representative" structures of "similar" design; as yet there is no equivalent wording for military aircraft. Although open to interpretation, the basis premise is that if you have designed and tested a similar component before, and if you can show an analysis method that gives an acceptable level of accuracy then you can clear a new "generic" component by analysis alone. Definition of structures within appropriate non-linear finite element (FE) codes is now well understood and developed, however, bird models vary considerably between workers. Differences in density, shape and aspect ratio are easy to see; differences in mechanical properties are not as transparent. Unless data to define analytical birds can be justified against viable sources, it is unlikely that certification of structures by analysis alone will be successful. Similar thoughts apply when considering substituting real birds for synthetic birds in testing. It is important that data used in defining bird properties has been taken from a justifiable source, not just assumed or fitted to test data. Previous workers (Ref. 2) have stated that “Scientists should aim to make the model fit the bird, not make the bird fit the model”.